PSI - Issue 66

Karolina Głowacka et al. / Procedia Structural Integrity 66 (2024) 108 – 121 Author name / Structural Integrity Procedia 00 (2025) 000–000

111

4

In the DIC analysis, two data filters were used - median surface filter and smoothing spline- to eliminate abrupt changes in values over time. The DIC analysis was carried out in the GOM Correlate program. In order to define the directions of deformation used in further analysis, in Figure 2, the directions of strains in the direction of the fiber orientation, as well as the length of the sample ε x and along the height of the sample ε y , were marked. The shear strain γ xy is defined in the plane defined by these directions. Table 1 compares the notation used in the program to the Voigt notation commonly used for composite materials.

Figure 2. Definition of individual strain directions used for analysis

Table 1. Notation of the analyzed deformations - comparison in Voigt notation and in the GOM Correlate program

ε � [-] ε � [-] γ �� [°]

ε � [-] ε � [-] ε � [rad]

Notation in GOM

Marking

Voigt notation

Strain along fibers Strain across fibers

Black arrows Grey arrows

Shear strain in the analyzed plane

-

3. Results Figures 3-9 show graphs of the relationship between force and displacement of the loading pin during bending. Depending on the selected sample size and support span, different ratios of normal to shear stresses occurred, which was the planned procedure. As a result, the samples were destroyed in other ways, which can be seen in the photos presented in these figures. Five samples of each type were analyzed, so five curves were presented, but one curve was bolded on each of the graphs and subjected to a broader analysis. In the case of repeatable sample behavior, one representative sample was analyzed for further study. In contrast, in the case of diversified sample behavior, a more comprehensive range of samples was presented in the photos. Each figure consists of several related drawings. First, a bending graph is presented, together with the characteristic stages of bending marked. Then, the subsequent photos show an image of the bent sample at a given moment (corresponding to the point on the graph). In the case of points marked with number 1, in each case a view of the sample in the frame was presented just before destruction. Then, the section where the fracture was supposed to occur is shown in a closer approximation, and the same section of the sample is shown after being analyzed using digital image correlation. Subsequent characteristic bending stages were not analyzed using digital image correlation because, in most cases, the sample was already so deformed that further analysis would lead to erroneous conclusions.